An efficient way of guaranteeing the survival of seedlings in areas under water stress

Afforestation and forest restoration in Mediterranean drylands frequently fail because summer drought creates a demographic bottleneck. In these water-limited ecosystems, the establishment of Quercus rotundifolia Lam., a late-successional species, is essential for rebuilding the tree layer and enabling community assembly. Ensuring early survival is critical for initiating ecosystem recovery under arid conditions. Global climate change reinforces the need to optimize resource use, particularly water, which is limiting across the Mediterranean Basin. Water scarcity reduces forest functioning and threatens ecosystem services such as carbon sequestration and productivity. The Nodriza device is a passive reservoir designed to capture precipitation and gradually deliver water to the root zone during dry periods, buffering seedlings against water stress. This study evaluates the efficiency of the Nodriza device in improving soil moisture, seedling performance and survival, and compares it with a commercial device (Groasis) in a reforestation using Quercus rotundifolia. Two field experiments with Q. rotundifolia seedlings were conducted in Mediterranean environments. Experiment A compared Nodriza with non-irrigated controls, while Experiment B compared Nodriza, Groasis and control treatments across a heterogenous soil. Soil moisture at 10–40 cm depth, microclimate, growth, and survival were monitored and linear mixed models and generalised linear models were used to analyse treatment effects. In Experiment A, Nodriza improved soil moisture at 20–40 cm depth, prolonging water availability into late summer and reducing water stress. Seedlings with the Nodriza showed greater height and basal diameter at the end of the experiment, and survival reached 87%, compared with 0% in the Control treatment. In Experiment B, Nodriza maintained higher soil moisture than Groasis and reduced internal temperatures. Nodriza seedlings showed growth and achieved 100% survival, exceeding Groasis (87%) and controls (70%). Overall, the Nodriza device improved soil-water availability, moderated microclimatic extremes and increased survival, making it a potential tool for plant restoration in Mediterranean drylands.

Recent studies have emphasised the severe and growing risks confronting areas such as the Mediterranean that are subject to accelerating climate change and endangered by its impact on water, ecosystems, food, health and security (Cramer et al., 2018; IPCC, 2023; European Environment Agency, 2024). Predictions for the period 2070–2100 in the Mediterranean indicate that mean temperatures will rise in summer and winter temperatures will be milder, and that precipitations will decline throughout the whole of the year (Ozturk et al., 2015; Cos et al., 2022).

Nevertheless, as a result of climate change the Mediterranean region may be affected by serious floods provoked by extreme rainfall and by higher temperatures (Zittis et al., 2021; Seker and Gumus, 2022). Such severe rainfall generally occurs at the end of summer and beginning of autumn but is not usually sufficient to replenish reservoirs or restore optimum water levels after a long dry summer period (Gaume et al., 2016). Neither is it adequate for re-establishing soil moisture levels since most of this type of rainfall is torrential (Inoubli et al., 2017). These erosive phenomena are aggravated by factors such as the type of soil and soil uses (Serpa et al., 2015). This scenario is particularly relevant in Mediterranean-climate regions in Africa, America, Australia and Europe, where successful establishment of newly planted forest seedlings largely depends on maintaining adequate soil-water availability during the first months after planting. Increasing drought intensity and higher evaporative demand under climate change reinforce the need for strategies that regulate water supply during this critical establishment phase (Pardos and Calama, 2022; Lázaro-González et al., 2023; Vizinho et al., 2023; García-Fayos et al., 2024; Homet et al., 2024).

Together, the overexploitation of water resources and the effects of climate change have led to a widespread decline in water reserves (Rodell et al., 2018; Scanlon et al., 2023). Some studies have proposed replacing high-water-demand crops with more drought-tolerant species and improving irrigation efficiency as key adaptation strategies under climate change (Smith et al., 2019; Bruno et al., 2023; Boser et al., 2024). However, regardless of the species selected, successful establishment remains a major challenge because newly planted seedlings are highly vulnerable to water deficit until they develop a sufficiently deep and functional root system capable of accessing soil moisture independently. Consequently, ensuring adequate water availability during this critical establishment phase is essential for the success of both agricultural and forest plantations.

Ecological restoration has emerged as a key global strategy to mitigate climate change and reverse land degradation, and it is promoted through major international initiatives such as the United Nations Decade on Ecosystem Restoration (2021–2030) (Aronson et al., 2020; Abhilash, 2021; Nelson et al., 2024). Forest restoration and afforestation efforts in Mediterranean regions are widely recognized as nature-based solutions that enhance carbon sequestration biodiversity conservation, and ecosystem resilience (Rey Benayas et al., 2009; Cunningham et al., 2015). Although Mediterranean woody species are naturally adapted to seasonal summer drought, planted nursery seedlings experience a critical establishment phase during which their root systems are not yet sufficiently developed to access deep soil-water reserves. Consequently, first-year mortality can be high, particularly under the increasingly frequent and intense droughts projected for the Mediterranean Basin under climate change (Giorgi and Lionello, 2008; IPCC, 2023; Andres et al., 2024). The use of nursery-grown seedlings increases restoration success, but substantial mortality may occur during the first year after planting (Bainbridge, 2012; Iakovoglou et al., 2020; Lund et al., 2024).

In Mediterranean woodland systems, Q. rotundifolia Lam. functions as a structural keystone species whose establishment determines the regeneration of the tree layer and the initiation of community-assembly processes. Its failure to establish under increasing drought conditions keeps degraded landscapes trapped in arrested succession states dominated by herbaceous layers, preventing the recovery of facilitative interactions, microclimatic buffering and zoochorous seed-dispersal networks (Montserrat Martí et al., 2008; Gavinet et al., 2015; Pardos and Calama, 2022). This primary abiotic filter is expected to intensify under current climate-change scenarios, which predict longer and more severe drought periods across the Mediterranean Basin (Giorgi and Lionello, 2008; IPCC, 2023).

Given these limitations, several eco-technologies have been developed to buffer early-stage drought stress and improve seedling establishment in semi-arid restoration contexts (Piñeiro et al., 2013). Devices that manage rainfall, regulate water delivery to the root zone and provide microclimatic protection have been increasingly adopted as complementary tools in afforestation projects. These devices generally have a deposit for storing rainwater, as well as in some cases mechanisms for gathering moisture from condensation when environmental temperatures are low. Each device employs a different mechanism for redirecting water to the plant roots, generally via capillary activity in and around the roots. However, many existing solutions still face limitations due to their inability to ensure deep-soil moisture availability during the peak of the summer drought.

In this context, we present the patented “Nodriza” device, a passive hydraulic system designed to secure seedling survival during the early stages of root development and to buffer the effects of erratic precipitation patterns. Unlike condensation-based protectors, this device functions as a controlled-release reservoir that mitigates the decline in soil moisture during the critical summer months. The aim of this study was to evaluate the Nodriza device in facilitating the establishment of Q. rotundifolia seedlings under field conditions in Mediterranean region a dry ombrotype and arenosol soils.

The specific objectives were: (i) to quantify the improvement in soil-moisture availability compared with the control and with alternative commercial devices independently; and (ii) to evaluate the impact on seedling survival and growth during the first year after planting, testing the hypothesis that the incorporation of these devices, which deliver water close to the seedling root system, effectively prevents the summer-drought bottleneck.

Description of the patented device Nodriza

The patented device with identity code ES2351840A1, Device for gathering rainwater for irrigation with protector (Figure 1). The device—capacity approx. 48 L and dimensions 48 × 43 × 33 cm (length, width, height)—is constructed from opaque polyethylene to prevent the grow of photosynthetic organism’s in the stored water that could obstruct its flow to the outside. Its top is sealed with a perforated cover that allows rainwater to enter (Figures 1 and Supplementary Figures S1A,B). At the base, the outlet pipe consists of a 20 mm low-density polyethylene tube with a dogleg into which the ceramic cone (red clay fired at 600 °C) is completely buried in the soil at a 20 cm approximately and in contact with the root system (Figures 1 and Supplementary Figures S1C,D). The porous cone allows water to be released only when the surrounding soil moisture decreases, thereby supplying water directly within the root zone.

Although the nominal storage capacity of the reservoir is approximately 48 L, its effective capacity under field conditions is about 35 L because the lid occupies part of the reservoir volume and the device is rarely installed on perfectly level ground. Water is released passively through the porous ceramic cone according to the hydraulic gradient between the reservoir and the surrounding soil. Under Mediterranean summer conditions, a reservoir containing between 30 and 35 L of water maintained elevated soil moisture in the root zone for an average of 56 days, corresponding to an approximate water supply of 0.53 L plant −1 day −1.

The device is also equipped with a simple structure providing shade and protection for the seedling (Supplementary Figure S1A) that consists of a metal frame covered by garden mesh to which a vertical roll of metal or plastic mesh (depending on the type of site in which the reforestation is being undertaken) is attached. If it is a cattle farm with herbivores, a perforated metal mesh secured to the soil with corrugated steel stakes must be put in place (Supplementary Figure S1A). The thickness and length of these stakes must be sufficient to prevent cattle from moving the cover if they attempt to graze the seedling.

All seedlings used in both Experiment A and Experiment B were 2-year-old Q. rotundifolia obtained from certified forest nurseries of the Junta de Andalucía (Lugar Nuevo, Jaén, Spain). Seedlings were grown in standard forest trays (≈300 cm 3) under conventional nursery practices. Prior to planting, individuals were sorted by height and root-collar diameter to ensure homogeneity among treatments. Individuals with signs of disease, malformation, or size outliers were discarded. Regarding the initial status, seedlings were planted between February and March depend on the experiment. Initial biometric measurements confirmed the homogeneity of the sample with the absence of significant differences between groups (p > 0.05).

Efficiency of the Nodriza in the field (experiment A)

The experiment A to test the patented Nodriza device was carried out in farm called “Las Hazas,” in the Villanueva de Córdoba village (Córdoba, Spain; 38°25′43″N 04° 35′49”W) between March 2012 and April 2013. A total of 41 western holm oak (Q. rotundifolia) seedlings, 2 years of age, were planted, of which 30 were equipped with the patent (Nodriza) and 11 without (Controls), at 10 meters between each (Supplementary Figure S2). Plant positioning within the plot was a fully randomised spatial arrangement.

The farm has an ecological agricultural certification and its stock—beef cattle, Iberian pigs and sheep—grazes extensively in wood pasture. It was necessary to adapt the Nodriza device to protect the seedlings from herbivory by placing an additional protective mesh inside the standard mesh. Sowing was performed with a tractor arm dug holes of approximately 0.5 m 3 in volume (SAME Treviglio, Italy). All planting was carried out on a single day (17 March 2012) and each sample plant was watered initially with 35 litres of water. At the same time, a tube for the probe used to measure the moisture percentage of the soil was buried as close as possible to each seedling. A PR2/4 probe (Delta-T Devices, UK; Supplementary Figure S1C) was used to measure the moisture content of the soil once a month three times per plant—but once every 2 weeks during the driest time of year (from June to August)—at depths of 10, 20, 30, and 40 cm from March 2012 to April 2013. At each sampling date, the PR2/4 probe was inserted into the single access tube of each plant, and three consecutive readings were taken, removing and reinserting the probe between measurements to reduce measurement error. Although measurements started immediately after installation to establish a complete temporal record from planting onwards, the treatment comparisons discussed in this study are primarily based on the summer period, by which time any disturbance associated with tube installation would be expected to have substantially diminished. In July 2012, 35 litres of water were added to each plant sample.

In addition, plant vigour was monitored: during the first month after planting the diameter of the base of the main plant stem was measured using a digital calliper (PCE Instruments, Albacete, Spain), and the maximum height of the stem was measured with a flexometer. These measurements were also carried out before the beginning of the dry season (June 2012) and at the end of the experiments (April 2013). Data regarding plant survival were recorded at the end of the experiment in April 2013.

Meteorological data were taken from the weather station in the village of Villanueva de Córdoba (Agencia Estatal de Meteorología - AEMET, 2024) (2012 to 2013) (Córdoba, Spain) (38°24′16.30″N, 4°36′20.93″W, 750 m a.s.l.) (Supplementary Figure S3A).

From a bioclimatic perspective, the study area lies in the domain of the Mediterranean macrobioclimate, Mediterranean Pluviestational Oceanic bioclimate, with a Lower Mesomediterranean thermotype and a Lower Dry ombrotype (Rivas-Martínez et al., 2017). The dominant vegetation corresponds to the Mesomediterranean Luso-extremadurense dry-subhumid acidic holm oak (Quercus rotundifolia) association: Pyro bourgaeanae-Querco rotundifoliae S. (Costa Pérez and Valle Tendero, 2004). This association is dominated by cleared wood pasture with large holm oaks and practically no understorey since it has been cleared to improve the grazing for the various types of cattle on the farm. The dominant plant species are Quercus rotundifolia, Pyrus bourgaeana Decne., Q. coccifera L., Phillyrea angustifolia L., Juniperus oxycedrus L. subsp. badia (H. Gay) Debeaux, Rhamnus alaternus L., Asparagus acutifolius L., Cistus salviifolius L., C. ladanifer L., Lavandula stoechas L., Erophaca baetica (L.) Boiss. subsp. baetica, Poa bulbosa L. and Trifolium subterraneum L. (Blanca et al., 2011).

The study area is situated on Los Pedroches Batholith, which runs WNW-ESE from the village Linares (Jaén, Spain) to the village La Serena (Badajoz, Spain) (Larrea et al., 2013). Its igneous-plutonic and phyllonian rocks give rise to shallow Cambisol soils on the metamorphosed rocks. Horizon A is shallow, around 10 cm in depth and with a sandy loamy texture. Horizon B is paler with a finer texture than A. Finally, there is a transitional horizon BC containing pieces of weathered rock at a depth of 20–50 cm. The pH slightly acidic, around 6–6.1 (Junta de Andalucía, 1999).

Efficiency of the commercially available device vs. the Nodriza device in the field (Experiment B)

The field experiment B aimed to compare the patented device (Nodriza) with the commercially available (Waterboxx Groasis ®, https://www.groasis.com/es), an internationally recognised patent (WO 2015/063243 A1). This experiment B was performed on a farm in the municipality of Bailén (Jaén, Andalusia, Spain; 38°07′02.7″N 3°47′08.7″W; Supplementary Figure S2). It lies on Mesozoic materials from the Triassic Buntsandstein consisting of conglomerates and red sandstones mixed with red clays (Mudarra Gómez et al., 1976).

This experiment B with the patented device involved planting 15 seedlings (2-years-old) equipped with the Nodriza, 15 with the Groasis and 10 with no irrigation device (Control) (Supplementary Figure S2). The experimental design was based on random blocks, with three treatments (Nodriza, Groasis, Control) and three heterogenous planting lines with a distance of 10 meters between lines and three meters between plants. Each of the three planting lines functioned as an independent block. Within each line, the three treatments were randomly assigned and repeated sequentially along the row, ensuring that every block contained one replicate of each treatment. This produced a balanced, fully randomized block design despite the linear spatial structure of the site.

Groasis device consists of a sealed deposit, with a capacity of around 15 L, and 48.5 cm in diameter and 23 cm in height (Empower nature, Waterboxx ®). It weighs approximately 1,400 g and is made from virgin polypropylene with an extra anti UV-radiation coating. The device is located at ground level and the plant is placed in its centre. Under arid conditions, the device helps condense the dew that forms when temperatures drop at night in these environments. The device supplies water to the soil around the plant via a series of capillary strands of textile, around 13–15 cm in length, located around the base of the plant.

Once again western holm oak (Q. rotundifolia) was chosen for planting in 0.5-m 3 holes, excavated with hoes, followed by a manual homogenization of the soil. All planting took place on 7 February 2013; the Nodriza and Groasis devices were filled and each Control plant watered with 30 L of water. In this experiment the deposits were not refilled in the dry season given that in spring 2013 (March–May) 289.8 mm of rain fell, almost half of the total annual rainfall according to data from the IFAPA-Mengíbar weather station (Junta de Andalucía, 2026).

A tube was buried next to the seedling roots to measure the moisture content in the soil at depths of 10, 20, 30, and 40 cm. These measurements were taken once a month among February 2013 to February 2014 but every 2 weeks in the dry season (from May to August 2013). At each sampling date, the PR2/4 probe was inserted, and three consecutive readings were taken, removing and reinserting the probe between measurements to reduce measurement error. Monitoring of plant vigour was assessed by measuring the basal diameter and height of the main stem, and survival rates. These measurements were taken after the initial planting and at the end of the experiment. In addition, the mean (T), maximum (Tmax) and minimum (Tmin) temperatures at the base of the plant and 1 m away of the devices were taken from eight plants with Groasis devices, eight with Nodriza and three Controls. The mean percentage (H), maximum (Hmax) and minimum (Hmin) of the relative humidity were taken on the plants. These measurements were taken immediately after planting and again at the end of the experiment. In addition, air temperature and relative humidity were measured immediately adjacent at the seedling base (hereafter seedling base) within the corresponding protection system (Nodriza, Groasis or the wire-mesh protector used for Control seedlings) and at a reference point located 1 m away (hereafter 1 m from seedling) using a portable digital thermo-hygrometer with an external probe (Essentials Digital TH-812E, Microtemp Electrics, Taiwan). The instrument simultaneously measured air temperature and relative humidity, while the external probe recorded temperature at the selected location. The thermo-hygrometer also stored the maximum and minimum values of both temperature and relative humidity in its internal memory. Measurements were obtained from eight seedlings fitted with Groasis devices, eight fitted with Nodriza devices and three control seedlings, each seedling representing one independent replicate. One measurement was taken per seedling during each sampling campaign, and all measurements were performed consecutively within the same daily time window to ensure comparable environmental conditions among treatments. Mean (T), maximum (Tmax) and minimum (Tmin) temperatures together with Mean (H), maximum (Hmax) and minimum (Hmin) relative humidity were recorded for each sampling point. Data regarding plant survival were recorded at the end of the experiment in February 2014.

Meteorological data were obtained from the weather station at IFAPA-Mengíbar (Jaén) (37°19′59″N 4°36′32″W, 293 m a.s.l.) (Supplementary Figure S3B). In bioclimatic terms, this area lies in the Mediterranean macrobioclimate, Pluviestational Oceanic bioclimate, with an Upper Thermo-Mediterranean thermotype and a Lower Dry ombrotype (Rivas-Martínez et al., 2017). The dominant vegetation in the area around the estate where the second experiment was carried out corresponds to the Mediterranean Betic dry-subhumid basic holm oak (Quercus rotundifolia) plant association: Paeonio coriaceae-Querco rotundifoliae S. thermophile facies with Pistacia lentiscus (Costa Pérez and Valle Tendero, 2004). The farm is surrounded by olive groves with patches of the region’s potential vegetation along field margins. The most frequent species are Quercus rotundifolia, Q. coccifera, Phillyrea angustifolia, Rhamnus alaternus, Pistacia lentiscus L., Aristolochia baetica L., Smilax aspera L., Tamus communis L., Asparagus acutifolius, Cistus salviifolius, Lavandula stoechas and other grassland plant species such as Brachypodium retusum (Pers.) P. Beauv., Dactylis glomerata L. subsp. hispanica (Roth) Nyman and Retama sphaerocarpa (L.) Boiss. (Blanca et al., 2011).

Statistical analyses

Data were analysed using R version 4.3.2 (R Core Team, 2026) and R Studio version 2026.06.0.242 (Posit team, 2026), with the packages glmmTMB (Brooks et al., 2017). Residuals were analysed for validation with the package DHARMa (Hartig, 2024). The car package (Fox and Weisberg, 2019) was used to test the significance of the fixed factors with Wald’s test, as well as; the multiple comparisons between factors were assessed using Tukey-adjusted pairwise contrasts computed with the emmeans package (Lenth, 2025).

The percentages of the soil moisture at 10, 20, 30 and 40 cm were analysed with a Linear Mixed Model (LMM) for the interaction of depth and treatment as the fix factor and sampling date and individual plant as the random factors, to account the temporal variation and for repeated measurements within each plant. The models were fitted using Restricted Maximum Likelihood (REML) (Pinheiro and Bates, 2000; Zuur et al., 2009). The height and the diameter of the base of the main stem was analysed with a LMM separately with the interaction between treatment and sampling date as a fixed factor, and the individual plant as a random factor, to account for repeated measures over time and within each plant. The mean, maximum and minimum seedling base temperature and 1 m from seedling were analysed separately with a LMM, with treatment as the fixed factor and sampling date and individual plant as the random factor, to account for the temporal variations and for repeated measures within each plant. Mean, maximum and minimum air humidity were analysed separately with a LMM, with treatment as the fixed factor and sampling date as the random factor, to account for the temporal variations and for repeated measures within each plant. Survival rate in each experiment was analysing with a GLM with binomial distribution and treatment as fixed factor.

Efficiency of the Nodriza in the field (experiment A)

The results indicate that the soil moisture content (% volume of water) is affected by the depth (LMM, χ 2 = 2594.09, d.f. = 3, p < 0.001) and treatment (LMM, χ 2 = 99.61, d.f. = 1, p < 0.001) and the interactions between both variables (LMM, χ 2 = 153.41, d.f. = 3, p < 0.001) (Figure 2, Supplementary Table S1).

Soil moisture at a depth of 10 cm was significantly higher in the Nodriza (7.01 ± 0.13%, mean ± se) than the Controls (6.07 ± 0.28%, mean ± se) (Figure 2, Supplementary Table S1). The same pattern is registered at depths of 20, 30 and 40 cm, the moisture content was significantly greater in plants with Nodriza (11.04 ± 0.15% at 20 cm depth; 10.92 ± 0.17% at 30 cm depth; 11.60 ± 0.19% at 40 cm depth, mean ± se) than the Controls (8.46 ± 0.28% at 20 cm depth; 8.17 ± 0.24% at 30 cm depth; 9.19 ± 0.28% at 40 cm depth, mean ± se) (Figure 2, Supplementary Table S1).

Seasonally, mean soil moisture at the analysed depths in April ranged from 4.62 to 6.57% in the Control and from 7.34 to 12.12% in the Nodriza treatment. During the summer drought, soil moisture progressively declined, reaching values from 0.19 to 3.33% in the Control and from 2.10 to 5.36% in the Nodriza treatment by the end of August (Figure 3). Following the autumn rainfall recharge, soil moisture increased substantially in both treatments, reaching from 10.36 to 13.54% in the Control and from 11.97 to 16.25% in the Nodriza treatment in October (Figure 3). These values remained below the estimated field-capacity range reported for the representative soil profile of the study area. During the summer period, soil moisture remained consistenty hgher in the Nodriza treatment at 20, 30, and 40 cm depth, indicating contrasting seasonal soil-water dynamics during the peak summer drought.

Plant vigour differed significantly among treatments. The height of the main stem showed significant differences in the treatment and sampling date interaction (LMM, χ 2 = 24.15, d.f. = 2, p < 0.001; Figure 4A, Supplementary Table S1); however, neither factor had a significant effect when analysed individually (LMM treatment, χ 2 = 3.12, d.f. = 1, p = 0.08; LMM sampling date, χ 2 = 0.96, d.f. = 2, p = 0.62). Plant with Nodriza (32.01 ± 1.94 cm, mean ± se) were taller than Control plants (24.59 ± 1.36 cm, mean ± se) in the final assessment. Basal stem diameter registered differences in treatment and sampling date interaction (LMM, χ 2 = 46.05, d.f. = 2, p < 0.001; Figure 4B, Supplementary Table S1), treatment (LMM, χ 2 = 9.43, d.f. = 2, p < 0.01) and sampling date (LMM, χ 2 = 25.19, d.f. = 2, p < 0.001). Nodriza (5.84 ± 0.26 mm, mean ± se) had a higher basal stem than Control (3.71 ± 0.11 mm, mean ± se) only at the final sampling.

We did not record a treatment effect on the survival probability (GLM, χ 2 = 0.00, d.f. = 1, p = 0.99). The probability estimated of survival marginal were 0.00 ± 0.00% (95% CI: 0–100%) in the Control, whereas it reached 86.7 ± 6.2% (95% CI: 69–95%) under the Nodriza device. Nodriza treatment showed a higher probability, but the control range is extremely wide due to the complete separation (all individual died). This prevents us from concluding that real differences exist.

Efficiency of the commercially available device vs. the Nodriza in the field (experiment B)

The results indicate that the soil moisture content (% volume of water) is affected by the depth (LMM, χ 2 = 1257.78, d.f. = 3, p < 0.001) and treatment (LMM, χ 2 = 84.82, d.f. = 1, p < 0.001) and the interactions between both variables (LMM, χ 2 = 90.00, d.f. = 3, p < 0.001) (Figure 5, Supplementary Table S1).

Soil moisture at a depth of 10 cm was higher in the Nodriza treatment (17.64 ± 0.31%, mean ± SE) followed by Groasis treatment (15.36 ± 0.30%, mean ± se), and Control (15.89 ± 0.42%, mean ± se) (Figure 5, Supplementary Table S1). Furthermore, at a depth of 20 cm, Nodriza recorded also the highest moisture followed by Groasis and Control the lowest moisture significantly different between the latter two (Figure 5, Supplementary Table S1). At a depth of 30 cm Nodriza plants (19.78 ± 0.29%, mean ± se) had greater soil moisture than those with the Control plants (18.43 ± 0.41%, mean ± se) and the Groasis device (17.26 ± 0.27%, mean ± se), with significant differences between them (Figure 5, Supplementary Table S1). Finally, at a depth of 40 cm there were also significant differences in the soil moisture, Groasis treatment (19.45 ± 0.27%, mean ± se) registered a lower moisture than Control (22.54 ± 0.36%, mean ± se) and the Nodriza (22.51 ± 0.29%, mean ± se); nevertheless, between the Nodriza and Control plants no significant differences were found (Figure 5).

Seasonally, soil moisture ranged from 10 and 20% at all analysed depths at the beginning of the experiment and increased to more than 30% by April (Figure 6). Thereafter, soil moisture progressively declined until approximately early October, after which it increased again but never reached the peak values recorded in April at any of the studied depths (Figure 6). During the summer drought, differences in soil moisture among treatments were consistently maintained across all depths, indicating contrasting seasonal soil-water dynamics during the most limiting part of the year (Figure 6).

Data for the growth of the main stem showed significant differences on the treatment and sampling date interaction (LMM, χ 2 = 18.71, d.f. = 2, p < 0.001), treatment (LMM, χ 2 = 13.26, d.f. = 2, p < 0.01) and sampling date (LMM, χ 2 = 5.72, d.f. = 1, p < 0.02) individually (Figure 7A, Supplementary Table S1). The differences were only at the end of the experiment, the plants with Nodriza (40.19 ± 3.01 cm, mean ± se) were higher than Control (21.35 ± 2.58 cm, mean ± se) and Groasis (34.22 ± 1.71 cm, mean ± se) (Figure 7A). However, the diameter of the base of the main stem had significant differences on the treatment and sampling date interaction (LMM, χ 2 = 29.62, d.f. = 2, p < 0.001), sampling date (LMM, χ 2 = 43.40, d.f. = 1, p < 0.001) but not in treatment (LMM, χ 2 = 2.88, d.f. = 2, p = 0.24) individually (Figure 7B, Supplementary Table S1). Plants with Nodriza (7.50 ± 0.44 cm, mean ± se) recorded a significantly greater diameter than Control plants (5.07 ± 0.55 cm, mean ± se) by the end of the experiment, but without differences with Groasis (6.25 ± 0.26 cm, mean ± se) (Figure 7B, Supplementary Table S1).

Mean air temperature measured at the seedling base differed significantly among treatments (LMM, χ 2 = 29.42, d.f. = 2, p < 0.001). The Nodriza plants recorded lower temperatures (23.24 ± 0.67 °C, mean ± se) than both the Control (24.83 ± 1.19 °C, mean ± se) and Groasis (24.54 ± 0.68 °C, mean ± se) treatments, whereas no significant differences were detected between the latter two treatments (Supplementary Figure S4A, Supplementary Table S1). Mean air temperature measured at the reference point located 1 m from the seedlings showed the Control (19.97 ± 1.09 °C, mean ± se) had the highest temperature followed by Groasis (19.54 ± 0.70 °C, mean ± se) and Nodriza (18.63 ± 0.64 °C, mean ± se) (LMM, χ 2 = 56.95, d.f. = 2, p < 0.001) (Supplementary Figure S4B, Supplementary Table S1). Maximum air temperature measured at the seedling base was significantly lower in the Nodriza treatment (33.74 ± 0.81 °C, mean ± se) than in the Groasis (36.38 ± 0.77 °C, mean ± se) and Control (38.51 ± 1.18 °C, mean ± se) treatments, with significant differences also detected between the latter two treatments (LMM, χ 2 = 72.30, d.f. = 2, p < 0.001) (Supplementary Figure S4C, Supplementary Table S1). Maximum air temperature at the reference point was also lower for the Nodriza treatment (21.52 ± 0.55 °C, mean ± se) than for both the Groasis (23.18 ± 0.63 °C, mean ± se) and Control (23.39 ± 1.09 °C, mean ± se) treatments, whereas no significant differences were detected between the latter two treatments (LMM, χ 2 = 60.04, d.f. = 2, p < 0.001) (Supplementary Figure S4D, Supplementary Table S1). Finally, minimum air temperature measured at the seedling base was significantly higher in the Groasis (7.07 ± 0.67 °C, mean ± se) than in both the Nodriza (4.99 ± 0.58 °C, mean ± se) and Control (3.72 ± 0.86 °C, mean ± se) treatments (LMM, χ 2 = 16.61, d.f. = 2, p < 0.001; Supplementary Figure S4E, Supplementary Table S1). Likewise, minimum air temperature at the reference point was significantly higher in the Groasis treatment (16.08 ± 0.82 °C, mean ± se) than in the Control (15.11 ± 1.34 °C, mean ± se) and the Nodriza (14.91 ± 0.76 °C, mean ± se), with significant differences also between Control and Nodriza (LMM, χ 2 = 59.96, d.f. = 2, p < 0.001) (Supplementary Figure S4F, Supplementary Table S1).

Seasonally, the mean air temperature measured within the protection system, immediately adjacent to the seedling, and at the external reference point located 1 m from each seedling, and at the external reference point located 1 m from each seedling was approximately 10 °C at the beginning of the experiment and increased to values above 30 °C during July, August and September, before declining to values similar to those recorded initially (Figures 8A,B). Maximum air temperature followed the same seasonal pattern, increasing from approximately between 20 and 30 °C at the beginning of the experiment to values exceeding 30 °C inside the protection device and 40 °C at the external reference point during summer (Figures 8C,D). Likewise, minimum air temperature followed a similar seasonal trend, with initial values below 0 °C inside the protection device and around 10 °C at the external reference point (Figures 8E,F).

The Control plants (41.89 ± 2.21%, mean ± se) had significantly lower air humidity content than the plants equipped with either the Nodriza (46.98 ± 1.62%, mean ± se) or Groasis (46.99 ± 1.66%, mean ± se) devices, although there were no such differences between the two treatments (LMM, χ 2 = 11.88, d.f. = 2, p < 0.05) (Figure 9A, Supplementary Table S1). Humidity percentages began at around 50% and decreased until October and reached a minimum of less than 25%. Thereafter, values began to rise to reach a maximum of over 75% by the end of the experiment (Figure S5A). Maximum moisture also registered differences among treatments (LMM, χ 2 = 9.13, d.f. = 2, p < 0.05) (Figure 9B, Supplementary Table S1), Control plants (93.22 ± 0.98%, mean ± se) had the higher maximum moisture than Groasis (90.26 ± 1.13%, mean ± se), without differences with Nodriza (91.77 ± 0.82%, mean ± se). The moisture percentages were only below 100% at two moments during the summer (Supplementary Figure S5B). Lastly, the minimum moisture was significantly higher in the Nodriza (26.90 ± 0.66%, mean ± se) than in the Control (22.86 ± 0.59%, mean ± se) plants but not between the Nodriza and Groasis plants (26.02 ± 0.69%, mean ± se) (LMM, χ 2 = 38.64, d.f. = 2, p < 0.001) (Figure 9C, Supplementary Table S1). No minimum moisture values above 25% were detected at any point during the experiment (Supplementary Figure S5C).

We did not register an effect of treatment on the probability of survival (GLM, χ 2 = 0.99, d.f. = 2, p = 0.608). The survival probability showed values of 70 ± 15% (95% CI: 38–90%) for the control, 87 ± 9% (95% CI: 59–97%) for Groasis, and 100 ± 0% (95% CI: 0–100%) for Nodriza.

Our results demonstrate that the Nodriza device consistently reduced water stress, extended favourable moisture availability well into the dry season, and markedly increased seedling survival and growth across two contrasting Mediterranean environments. This pattern indicates that the device effectively helps seedlings overcome the early-stage demographic bottleneck imposed by summer drought, which is the principal limitation to afforestation success in semi-arid Mediterranean systems.

By ensuring the early survival of Q. rotundifolia, the Nodriza device effectively helps seedlings overcome the primary abiotic filter imposed by summer drought, which is widely recognized as the main demographic bottleneck in Mediterranean afforestation. Once seedlings surpass this initial mortality peak, they begin to modify the microenvironment through shading, soil–root interactions and litter inputs, generating small recruitment foci that act as nuclei for subsequent colonization. These early-established individuals facilitate the arrival of shrub and herbaceous species, including zoochorous taxa dispersed by birds and mammals, initiating nucleation and community-assembly processes that are otherwise arrested in degraded Mediterranean landscapes. In this sense, improving the establishment of Q. rotundifolia is not a single-species objective but the ecological gateway that enables the recovery of structural complexity and long-term ecosystem resilience.

Although Experiments A and B were conducted in different years, soil types and seasonal rainfall patterns, this heterogeneity strengthens rather than weakens the interpretation of device performance. The Nodriza system consistently maintained higher soil moisture and improved seedling survival across contrasting environmental conditions, showing that positive effects of the device were observed under two independent Mediterranean field conditions. Instead of acting as confounding noise, the variation between sites provides complementary validation, demonstrating that the device performs reliably under multiple Mediterranean scenarios typical of operational restoration projects.

The survival of any plant in a reforestation project depends on it being kept well-watered since the plant mass of the repopulated plants has a poorer capacity to derive moisture from the soil that surrounds it (Grossnickle, 2005; Villar-Salvador et al., 2012; Hawkins et al., 2014; Lázaro-González et al., 2023). Villar-Salvador et al. (2012) have shown that the survival of Quercus rotundifolia plants is related specifically to the development of its roots during the dry season. It has been demonstrated that if a plant absorbs less moisture than the amount it transpires it enters into water stress and physiological ‘shock’ or post-planting syndrome (Kaushal and Aussenac, 1989; Burdett, 1990). It is highly likely that this caused the development of the seedling roots in experiment A to be so slow and insufficient for withstanding the summer drought: this eventually led to the mortality of the plants that were not equipped with a patented device during the experiment A.

In experiment A the soil moisture content at all depths in the seedlings with the Nodriza device was always higher than in the Control plants until September, month in which heavy rain fell (104.4 mm). The generally low rainfall in spring 2012 compared to spring 2013 ensured that the soil to a depth of 10 cm was completely dry, even in the plants with the Nodriza device. Nevertheless, the Control plants suffered a high level of water stress, and all had died before the rains returned in September 2012. The marked increase in soil moisture observed during autumn 2012 in Experiment A coincided with an exceptionally wet autumn (348.4 mm accumulated between September and November), highlighting the importance of seasonal rainfall distribution in controlling soil-water recharge independently of treatment effects.

Throughout the experiment B, the evolution of the soil moisture shows that the Nodriza device conserves most humidity at all four soil depths from the end of spring onwards. This is the critical period for plant growth given the high temperatures and lack of rain. Aside from supplying water to plants derived from rainfall, the Groasis device is also able to store water captured from condensation (Groasis waterboxx, 2024). Yet, in the study area this type of water condensation does not occur. However, it does possess a diffuser that supplies water to the surface of the soil but not to the main part of the plant’s root system. For this reason, at a depth of 40 cm the Groasis plants had less soil moisture. At a depth of 30 cm, the moisture content is higher to Nodriza plants because it has a more efficient system of transporting the water with the plant roots. This in turn means that the period of both water shortage and water stress for plants is shorter with the Nodriza device. This characteristic of the Groasis device was noted by Mercé Arévalo (2013), who recommend that its water supply system be modified to allow for a better and more effective supply of water to the soil.

Plant vigour and growth was very similar for height and basal diameter of the main stem, a logical finding given the direct correlation between these two values. During the experiment A plants with no significant differences in both these variables were used in the Nodriza and Control groups. However, a year later these variables had changed significantly since the Nodriza plants had grown more than the Control plants. In all reforestation projects there is a period in which there must be a good hydraulic connection between the soil and plant roots so that the plant can develop and resist the harsh environmental conditions (Villar Salvador and Oliet Palá, 2021). Once summer had finished, all Control plants had died but only 10% of Nodriza plants were dead, which suggests that the repopulation had begun to become established.

In the experiment B, the initial height of the main stem in Nodriza plants was less than in the other treatments but by the end of the experiment the plants that had grown most were the Nodriza plants. During this experiment, spring was reasonably wet; the high summer temperatures extended into October and almost no rain fell until December. Although rooting probably occurred since the planting took place a month before the generous spring rains began, the aridity of the long summer undoubtedly caused a number of plants to die.

Likewise, the internal temperatures of both devices were one or two degrees cooler than the maximum temperatures during the most arid and hottest part of the year. However, between December and January the maximum temperatures inside both devices were lower than those of the Control. In terms of summer maximums, the Nodriza seems to be best equipped to handle extreme summer temperatures, even if the Groasis device provides best protection against winter minimums due to the greater degree of protection it affords the plants.

Unlike the Groasis Waterboxx®, which almost completely encloses the seedling, leaving only a central opening for shoot emergence, the Nodriza device incorporates a perforated plastic mesh cylinder that protects seedlings from herbivory while allowing continuous air exchange. An east-facing opaque cover additionally provides partial shading during periods of maximum solar radiation. Protective shelters are known to modify the microenvironment surrounding seedlings by reducing radiation load, buffering temperature extremes and influencing evaporative demand (Dupraz and Bergez, 1999; del Campo García et al., 2008; Navarro-Cerrillo et al., 2009; Puértolas et al., 2010; Navarro et al., 2024). Although these microclimatic effects may have contributed to reducing surface evaporation, they are unlikely to fully explain the higher soil moisture observed in our study. Notably, control seedlings were also protected against herbivory using wire-mesh guards, thereby minimizing potential bias associated with the presence of a protective structure. Although the protective structure of the Nodriza may contribute to microclimatic buffering by reducing radiation load and evaporation demand, the present experimental design does not allow its independent contribution to be separated from that of the passive hydraulic reservoir. Therefore, we cannot exclude a contribution of the protective structure to the observed differences in soil-water availability, but attributing the differences between Nodriza and Groasis specifically to the mesh would be speculative. The depth-dependent pattern of soil moisture provides additional evidence that the observed response cannot be explained solely by a surface microclimatic effect. At 10 and 20 cm, soil moisture was higher in Nodriza than in both Groasis and Control, whereas at 30 cm all three treatments differed significantly, following the order Nodriza > Control > Groasis. At 40 cm, Nodriza and Control did not differ significantly, while both treatments showed significantly higher soil moisture than Groasis. Thus, the observed pattern cannot be attributed exclusively to the protective mesh or to the hydraulic component alone. Rather, the results are consistent with a combined effect of passive water supply and microclimatic buffering, although the relative contribution of these mechanisms would require a factorial experiment specifically designed to separate them.

Although a certain consensus exists that the effects of drought and wildfires will be felt worldwide, it is nevertheless still difficult to predict with any degree of certainty how particular geographical areas will be affected by drought and/or forest fires (Aguilera et al., 2020). Nodriza seems to be best equipped to handle extreme summer temperatures, even if the Groasis device provides best protection against winter minimums due to the greater degree of protection it affords the plants.

From an operational perspective, the small plot size and limited number of plants restrict direct extrapolation to landscape-scale restoration. However, the magnitude of the treatment effects observed in survival and growth suggests that the technology merits evaluation in larger, operational afforestation projects. Under high drought stress, tools that enhance establishment can substantially reduce replanting needs and increase the cost-effectiveness of restoration interventions.

The transition from experimental plots to landscape-scale restoration requires a strategic approach. Given the unit cost and logistical requirements of these autonomous watering devices, their indiscriminate application over large areas may not be economically feasible. Instead, we propose their implementation within a precision-restoration framework (Suding, 2011; Keenan et al., 2024).

The device may be particularly suitable for nucleation strategies (Corbin and Holl, 2012; Bechara et al., 2016; Piaia et al., 2020; Holl et al., 2024), in which restoration efforts are concentrated in small, high-density planting islands or resource islands (Zahawi et al., 2013; Hulvey et al., 2017; Procknow et al., 2023). This approach, known as applied nucleation, has proven to be an efficient and cost-effective alternative to traditional planting, accelerating the recovery of ecosystem structure and function even when only a limited fraction of the surface is initially treated. By ensuring the establishment of Mediterranean tree species such as Quercus rotundifolia within these nuclei through the use of the Nodriza device, managers can create biological focal points that later expand through natural facilitation and different types of zoochorous dispersal, rather than attempting large-scale afforestation that often results in high mortality.

Likewise, the device is especially recommended for recalcitrant microsites (areas with south-facing slopes, skeletal soils or historical erosion) where traditional planting methods systematically fail (Adams et al., 2017). In these critical points, the higher per-plant investment is justified by the increased likelihood of success under stochastic episodes of prolonged drought. From a cost–benefit perspective, the initial investment in hydraulic devices must be weighed against the operational costs of failure. Conventional reforestation in Mediterranean drylands often incurs hidden costs derived from mortality rates exceeding 50%, which force repeated beating-up operations and extended maintenance (Oliet and Jacobs, 2012).

These autonomous irrigation reservoirs involve a higher initial cost than simple standard tree shelters, whose efficiency has already been questioned (Navarro-Cerrillo et al., 2009). Nevertheless, our results with the Nodriza device (up to 100% survival) suggest a shift in the cost model: moving from “low-cost/high-mortality” inputs to “high-investment/high-survival” systems. For large restoration areas, this implies a reduction in long-term expenses related to replanting labour, transportation and nursery-stock acquisition. The device may become cost-efficient when costs are calculated per surviving tree over a 3- to 5-year horizon, although this statement should be confirmed with specific long-term economic studies.

Cost-efficiency considerations. Although the implementation of the Nodriza device entails a higher initial investment than traditional planting, preventing first-year mortality may offset these costs by reducing the need for replacement planting (“beating-up”) and maintenance visits in semi-arid regions. The objective of this study is not to evaluate market competitiveness, but to demonstrate that, in scenarios of high abiotic stress, survival-assurance technologies can improve long-term restoration outcomes and therefore represent an investment in restoration success rather than an added operational cost.

Our study demonstrates clear improvements in soil moisture, seedling growth and survival during the establishment phase. However, several practical aspects deserve further investigation before large-scale implementation. These include the long-term durability of the materials under field conditions, their potential environmental impacts after prolonged deployment, the logistics and costs associated with device retrieval or replacement, and comprehensive life-cycle and economic assessments. Addressing these issues will help define the most appropriate restoration contexts in which the Nodriza device can be efficiently applied.

The Nodriza device proved to be an effective passive hydraulic tool for improving the establishment of Q. rotundifolia in Mediterranean dryland. By maintaining higher soil moisture during summer drought and moderating temperature extremes around seedlings, the device enhanced growth and survival compared with both untreated controls and, in several cases, with the commercially available alternative. These results show that Nodriza helps seedlings overcome the main abiotic barrier limiting afforestation success in such environments.

Device performance remained consistent across two contrasting sites and climatic years, indicating functional robustness under typical Mediterranean conditions. Although broader testing is still needed for large-scale application, the strength of the observed effects supports its further evaluation within precision-restoration strategies, especially in applied nucleation or microsites where conventional planting often fails. Although the Nodriza device involves a higher initial investment than conventional planting, its substantially higher first-year survival suggests that it may reduce long-term restoration costs associated with replacement planting and repeated maintenance. Nevertheless, formal cost-effectiveness analyses incorporating manufacturing, installation, maintenance and life-cycle costs are required before economic conclusions can be drawn.

Overall, the Nodriza system is a promising tool for securing the establishment of structural Mediterranean tree species in highly water-limited environments and for initiating the recovery of woodland structure and function.